Sudden death from cardiac causes, often due to heart rhythm abnormalities such as ventricular fibrillation (VF) and ventricular tachycardia (VT), claims approximately 225,000 persons annually in the United States (1999 Heart and Stroke Statistical Update—Dallas: American Heart Association, 1998 or approximately one per 1,000 population per year.
Nationwide, only two to five percent of those who suffer an out-of-hospital cardiac arrest are saved (New England Journal of Medicine 2000; Vol. 343: Pgs. 1259-1260). In grid-locked cities like New York City, analysis of the Emergency Medical Service/Fire Department data has shown that less than 1% survive the ordeal. Because of the abysmal survival rate, the first arrest is almost always the last.
Cardiac arrest is the abrupt cessation of the heart's mechanical function leading to loss of consciousness and the rapidly progressing sequence of heart and brain deterioration, irreversible heart and brain damage, and death. This cessation of mechanical function during an arrest is often caused by a sudden chaotic deterioration in the heart rhythm referred to as ventricular fibrillation (VF) or by the sudden onset of a very rapid and mechanically ineffectual rhythm called ventricular tachycardia (VT). In either case, the heart's normal function can usually be restored by a prompt and properly administered electrical shock to the chest, generally referred to as defibrillation, or by application of synchronized electrical pacing signals.
It has been estimated that once an arrest has occurred the mortality rate increases by 10% per minute until definitive therapy commences. If treatment has not yet commenced ten minutes into a cardiac arrest, there is little likelihood of recovery. The evidence that the response time has a profound influence on the rate of successful resuscitation comes from two types of analysis: a) comparisons among studies in which the response times were different; and b) comparisons within studies, where the results of a short response time are compared to a longer response time, within the same population. The discussion of comparisons among studies follows.
Data from New York City and Chicago illustrate the poor results associated with a long response time. There are an estimated 7,000 incidents of cardiac arrest each year in New York City alone. Emergency Medical Service figures during a five year period—1994-1999—revealed that a total of only 168 patients were successfully resuscitated. The average EMS response time, from the time the service was summoned, was reported to be seven minutes. Based on the 10% per minute mortality estimate, one would expect a 70% mortality for a 7 minute response time. The EMS survival figures are much worse than would be expected based on the ten percent per minute mortality estimate. The reason may be that the time from the event onset until actual defibrillation is substantially longer than the 7 minute EMS dispatch and transit interval. This prolongation includes delays: a) from the moment of onset of arrest until EMS is called (referred to below as the “pre-call” interval); and b) from the time of EMS arrival until life-saving therapy is begun (referred to below as the “pre-shock” interval). Even if the sum of these delays is only three minutes, the expected resuscitation rate plunges to a negligible value, and is consistent with the observed very low survival rate. In Chicago (population over 3 million [linear interpolation of 1980 and 1990 census data]; area 228 square miles) the reported results, though slightly better than in New York City, are poor nevertheless. Becker et al. report that 91.4% of patients were dead on arrival at the hospital; 6.8% died in the hospital and 1.8% were discharged alive (New England Journal of Medicine 1993; Vol. 329: Pgs. 600-606). The response time, defined as the interval from the 911 call to the arrival of the ambulance (referred to below as the “call-to-arrival” interval) ranged from 1 to 22 minutes; the median was 6 minutes.
Improved response times are obtained by so called “code teams” in a hospital setting. An analysis of data from Kaye et al. showed that 9.5% of 210 hospitalized patients who suffered a cardiac arrest survived to leave the hospital (Circulation 1999; Vol. 100: Abstract 1645, page I-314). Eisenberg indicates that survival rates for inpatient cardiac arrests have been reported to range from 0 to 29% (New England Journal of Medicine 2001; Vol. 344: Pgs. 1304-1313). The higher survival rates are attributable to faster reaction times in the hospital environment.
Still better results are reported from King County, Wash., where the emergency response system is unusually sophisticated. Sweeny describes a 30% survival to discharge figure (Annals of Emergency Medicine 1998; Vol. 31: Pgs. 234-240) for this group. Similar results are reported from Rochester, Minn. (1993 population 76,865; area 32.6 square miles); White et al. report the results of 158 cardiac arrests between 1990 and 1995. An analysis of their data, in which the survival to hospital discharge is calculated for all arrest victims shows the value to be 26%. They make a more meticulous effort than some prior investigators to define and measure their response time, and they used the “call-to-shock” interval. Analysis of their data shows the value to be 6.0 minutes. (The aforementioned Chicago data was based on a 6 minute “call-to-arrival” interval. The call-to-shock interval is the sum of the call-to-arrival interval plus the pre-shock interval.)
Gambling casinos are an ideal locale for the analysis of the relationship between response time and outcome for two reasons: a) because of the ultra-high level of scrutiny, the time of onset of arrest can be known accurately; and b) because of the high level of stress for some patrons, the arrest rate is enhanced. Valenzuela et al. report on a program in which casino security officers where trained and equipped to rapidly defibrillate arrest victims (New England Journal of Medicine 2000; Vol. 343: Pgs. 1206-1209). There were 148 patients who suffered a casino cardiac arrest in 10 casinos in Nevada and Mississippi during the years 1997-1999. Fifty six victims (38%) were resuscitated and survived to hospital discharge. The interval from “collapse-to-shock” was 4.4 minutes. It represents a very short response time since it is the sum of: (a) the call-to-arrival interval (used in the Chicago study), plus (b) the pre-shock interval, plus (c) the pre-call interval. (The 6 minute Minnesota result was based on call-to-shock, i.e. (a) plus (b).) The interval from collapse to paramedic arrival was 9.8 minutes, which (without early defibrillation by the security officers) would be expected to result in a negligible survival rate.
The two keys to high arrest survival rates are a short interval from onset of arrest until provision of defibrillator shock, and the presence or rapid arrival of expert medical personnel at the arrest site. The compelling nature of the relationship between response time and success rate is further demonstrated by comparisons within studies.
The Rochester study divided victims with ventricular fibrillation (84 of the 158 arrests) into two groups: one group resuscitated by the police, whose call-to-shock time was a mean of 5.6 minutes, and one resuscitated by paramedics, whose call-to-shock time was 6.3 minutes. The fraction of a minute difference in arrival time impacted the survival rate. In the police group, 58% survived to hospital discharge; in the paramedic group with the slightly delayed arrival time, 43% survived to hospital discharge. (These survival values are larger than the previously mentioned 26% because that value included other arrest victims [i.e., victims with asystole and pulseless electrical activity] who are far less likely to be resuscitated than victims with VF.)
The casino data also gives firm support to the relationship between quick shock and hospital survival. Again, looking only at the VF victims, of those who received their first shock in less than three minutes, 74% (26 of 35) survived to hospital discharge, whereas only 49% (27 of 55) who were shocked after three minutes survived to hospital discharge.
The best results, in terms of resuscitation rate, occur in the hospital cardiac electrophysiology testing laboratory. Here, during the conduct of arrhythmia evaluations (referred to as electrophysiologic studies) in high risk patients, life-threatening VT and VF are frequently encountered. However, because of the presence of trained highly experienced arrhythmia physicians and nurses at the procedure, and because of the very short response time (time from onset of VF until time of shock is usually less than 15 seconds), the resuscitation rate is significantly greater than 99%.
The near 100% resuscitation rate in the electrophysiology laboratory represents an ideal that is not likely to be reproduced outside of the laboratory because:
a) the arrests in such a laboratory are all due to VT and VF, not asystole or pulseless electrical activity;
b) these arrests are artificially induced; hence, they are primary electrical disturbances, not electrical disturbances secondary to some other process, such as the sudden blockage of a coronary artery;
c) the response time is extremely short; and
d) doctors and nurses specializing in heart rhythm treatment are present at the procedure.
Nevertheless, the electrophysiology laboratory data does show there is nothing about VF per se that implies its irreversibility. As long as the response time is very short and the VF is not secondary to a sudden catastrophic structural problem (such as the abrupt blockage of an artery within the heart), we can expect a very high success rate. This concept is supported by the resuscitation results in patients with implantable cardioverter-defibrillators, which automatically detect and terminate VT or VF. Their response times and success rates, for spontaneously occurring VT and VF are comparable to those of the electrophysiology laboratory. The high success rate for very prompt termination of VF was confirmed in an entirely different setting. Page et al. report initial termination of VF in 13 of 15 (87%) patients who were treated as part of an effort to provide commercial airliners with defibrillators (New England Journal of Medicine 2000; Vol. 343: Pgs. 1210-1216).
As the time until shock increases, two types of events seem to occur which markedly reduce the chance of success. First, there is evidence that the longer the response time, the smaller is the fraction of patients actually found to have ventricular fibrillation. In other words, it may well be that in the moments immediately after an out-of-hospital arrest, the fraction of patients with VT or VF among all arrest victims is high; and that as the minutes go by, that fraction decreases. In the casino study, with its very short response time, 71% of victims had VF. In the Rochester study, with its intermediate response time, 53% had VF. In the Chicago study, with its long response time (since the 6 minute reported response time included only the call-to-arrival component), 22% had VF (calculated from their data). The fraction of victims with VF is important because among arrest causes, VF is far more likely to be treatable than either asystole or pulseless electrical activity. (In the Minnesota study there were 74 non-VF victims; and in the casino study there were 43; none of these non-VF victims were resuscitated.)
The second deleterious event which occurs very quickly during VF is the onset of irreversible mechanical damage to the heart muscle. Once such damage occurs, the chance of survival to hospital discharge plummets. The Minnesota study analyzed this by looking at a predictor of survival that they called “ROSC,” restoration of spontaneous circulation. ROSC was defined as present when either no cardiopulmonary resuscitation (CPR), or less than one minute of CPR was required. Victims with ROSC also did not require any medication to support their blood pressure. ROSC was a powerful predictor of survival to hospital discharge. Twenty seven of 28 victims (96%) with ROSC survived to hospital discharge; 14 of 56 without ROSC survived to discharge. The police with their 0.7 minute earlier arrival time had a much higher ROSC rate (42%) than did the paramedics (28%).
Automatic external defibrillators or AEDs were used by the police in the Minnesota study. They were used in the airline study and by the casino security officers. These AEDs are intended for use by minimally trained personnel. AED electrodes, which must be properly placed in contact with a cardiac victim's chest wall, allow the device to analyze the electrocardiogram (ECG) signals of a cardiac victim. Based on the ECG signal information which it receives, the AED automatically applies a high defibrillation voltage to these electrodes when its algorithm detects VT or VF. The decision to shock or not to shock, and the magnitude of the voltage application, are determined by circuitry within the device.
A number of systems are known which provide automatic external defibrillation. Equipment of this type is currently distributed by Medtronic Physio-Control, Philips and Cardiac Science, and may be purchased at a cost of about $2,500-$3,000 per unit. This equipment is now intended to be made available at places such as government buildings, casinos, airports, office buildings and sports arenas, and to be carried upon public modes of transportation such as commercial airliners. There is an increasing effort to have them carried aboard police cruisers.
The advantage of AEDs is that they allow a decreased response time by empowering non-medical people who can arrive sooner than paramedics to treat a cardiac arrest. In the Minnesota study, most police cars carried AEDs; they arrived sooner than the paramedics; and they had better results. In the casino study, cited above, security officers defibrillated with AEDs at a mean of 4.4 minutes after victim collapse; the paramedics arrived at a mean of 9.8 minutes after collapse.
On an airliner, the chance of arrest survival without on-board treatment is nil. The cardiac arrest survival rate in Boston increased from 16% to 24% when the number of AEDs was increased from 85 to 185 and all 1650 firefighters were trained to use AEDs and to perform cardiopulmonary resuscitation (Circulation 1998; Vol. 97: Pgs. 1321-1324).
Although AEDs have improved survival by decreasing response time they cannot be considered to be the ultimate solution because they lack certain important advantages that a highly trained medical professional possesses. The transfer of a responsibility, which traditionally lies within the domain of the medical profession, to the AED and its operator results in or fails to completely address seven classes of potential problems:
a) the limitation of proper AED performance to conditions addressed by its algorithm;
b) the necessity of assuring the proper electrical interface between the AED and the victim;
c) the persistence of delays not entirely circumvented by the AED;
d) the problem of CPR administration;
e) the problem of potential AED malfunction;
f) the aggravation of problems (b) through (e) in the event that the user of the AED is untrained or inadequately trained; and
g) the potential aggravation of any of problems (b) through (f) by the absence of a highly experienced professional taking charge of the emergency scene.
The seven classes of problems (a) through (g) will now be addressed and discussed in detail.
First, an AED relies on its internal artificial intelligence to make a decision about whether to provide a proper high voltage response for termination of a life threatening heart rhythm. The device must be programmed to anticipate as many situations as possible, and it must be programmed to function appropriately during each of those situations, in order for the automated response to have the intended effect of resuscitating the cardiac victim. The use of the device thus decouples the victim's treatment from the intelligence and judgment of a medical professional who normally administers external defibrillation. This usurpation of the medical professional's role by a machine and its minimally trained or untrained operator may, at times, result in incorrect or delayed responses in just those critical moments which can make the difference between life and death.
AEDs, no matter how complex their algorithms are or will become, are not able to perform properly under conditions which are not explicitly addressed by their algorithms. For example, Kanz et al. (Circulation 1999; Vol. 100: Abstract 1641, Page I-313) showed that AED-based rhythm diagnosis was often incorrect in the setting of substantial external electromagnetic interference. When 12 units were evaluated in railway stations, sensitivity ranged from 80 to 100%, and specificity ranged from 38 to 100%. In power stations, the performance was even worse, with both sensitivity and specificity ranging from 0 to 100%.
Other important issues which may not be addressed by an algorithm include the management of the victim who is fully or partially conscious with a tachycardia, and the management of a victim with an implantable cardioverter-defibrillator or ICD. Still other issues beyond the scope of current algorithms involve advanced management considerations such as post-defibrillation treatment.
Sophisticated EMTs will not benefit by carrying AEDs since they would be likely to know far more than the information on which an AED algorithm is based. However, a defibrillator device which provides the EMT with an immediately available medical expert consultant could improve arrest outcome.
The second limitation of AEDs relates to electrode positioning. It is known that correct defibrillator pad positioning and application is very important for successful defibrillation. Errors in positioning and poor electrical contact are not uncommon among inexperienced operators. AEDs do not actively guide the user in appropriate pad placement and application (other than by the provision of a diagram). Nor can they detect or correct for inappropriate positioning and application, once it has occurred. A defibrillator device which could provide such guidance would be highly desirable. Although highly sophisticated electronic means could provide such guidance, a human observer with means for observing pad placement could easily accomplish this.
Third, even partially trained AED users can not be expected to match the skills of a highly trained medical professional. For example, the casino study showed that 0.9 minutes elapsed from the time of defibrillator attachment until the time of first shock. A medical professional could accomplish this action in a fraction of this time. A defibrillator device which lets a remote ultra-sophisticated medical professional deliver the shock would therefore save time, when compared with the casino scenario.
Fourth, an AED does not coach an untrained bystander in the performance of CPR. Although CPR is not required in arrests of short duration, the need for it increases as the arrest duration increases. CPR was administered to some patients in the Minnesota study and in the casino study. The improved results in Boston were concomitant with not only an increase in available AEDs but with firefighter CPR training as well. Ewy, in discussing successes with a limited form of CPR which involves chest compression without ventilation, points out that rapid defibrillation and bystander initiated CPR are the major determinants of survival of a VF arrest (New England Journal of Medicine 2000; Vol. 342: Pgs. 1599-1560). A defibrillator device which could provide CPR instruction and guidance would be very advantageous. Although instruction prior to CPR could be automated, the processes of guidance during CPR and of suggesting corrective maneuvers during CPR, are far more easily accomplished by a human coach than by an algorithmic one.
Fifth, occasional malfunction of any electrical device is inevitable. Sweeny (cited above) noted seven instances of apparent AED malfunction out of 260 uses. It is far more likely that a medical professional who (i) has expert knowledge of a sophisticated defibrillator device and its backup systems, and (ii) constantly monitors the functioning of the defibrillator device during its operation, would be able to work around a device malfunction. (In the Sweeney study, use of AEDs in a Charlotte, N.C. EMS program did not result in outcome improvement.)
Sixth, in each of the reports cited herein, in which AEDs were used, their use was by a trained operator. It is inevitable that an untrained user will perform less accurately and take more time to do so. However, given that: (i) the ideal response time after an arrest would be even less than that during the casino study (in which victims were essentially under constant observation), and that (ii) the police or fire department response time is unlikely to ever be shorter than the casino response time; then the only likelihood of achieving the requisite ultra-short response time is by having a device that can be used by an entirely untrained bystander. Such a device would have to be more user-friendly than an AED. Such a device would have to be capable of both: (i) defibrillation, and (ii) closely linking an untrained bystander with an expert medical professional who could guide him through every aspect of the resuscitation process. AEDs do not meet this requirement. Indeed, the AEDs which have been installed to date typically display a warning to the potential user that the device is intended for use by trained personnel only.
Finally, since the aforementioned ideal external defibrillator device will require some level of participation by a human enabler (that is, a non-medically trained person who is available to use the device to defibrillate a victim of cardiac arrest), the creation of an environment in which the enabler functions optimally is critical. AEDs cannot address the anxiety or reluctance of an individual operator and may, in fact contribute to these. The cumulative effect of such feelings among a group of bystanders, may contribute to the chaos and pandemonium which not infrequently accompany a cardiac arrest. On the other hand, the voice of an experienced medical professional, taking charge, providing instructions, and making decisions, is often a great source of reassurance and stability, giving the assurance of proper conduct. AEDs cannot provide this human element.
Clearly, once a patient has suffered and survived an initial, lifethreatening cardiac event, that patient must be monitored closely so that the proper treatment may be brought to bear on an emergency basis. The U.S. Pat. No. 5,544,661 to Davis et al. discloses a patient monitoring system which includes a portable device, attached to a patient, and a central station. The portable device includes an ECG and a photo-plethysmograph connected to the patient, and an arrhythmia analysis circuit which includes an expert system for determining whether pre-established critical parameters have been exceeded. The portable device also includes a wireless wide area communication circuit for automatically contacting the central station via a public cellular telephone network when the expert system determines that assistance is needed. When the central station is contacted, the patient's ECG waveforms, measurements and trends are sent to the central monitoring station and a two-way voice channel between the patient and the central station is automatically opened. The central station includes a computerized facility from which a clinician can observe both the real time data being sent from the patient and the patient's historical records. The clinician can talk to the patient through the two-way voice channel and can also activate therapeutic devices attached to the patient such as an external defibrillator, a pacer or an automatic drug infusion device.
Similarly, the U.S. Pat. No. 5,564,429 to Born et al. discloses a cardiorespiratory alert system which comprises a patient unit, a base station and a remote unit. In a hospital configuration, several patient units can communicate with the base station, which is located centrally, for example at a nurse's station. In a home configuration, the base station can reside near the patient's bedside. In both cases, the communication between the patient unit and the base station is by way of radio telemetry. The base stations are designed to communicate with a remote unit, either by radio telemetry or by use of commercial telephone lines.
The system provides alerts to the remote unit when life-threatening conditions are detected in a patient, yet it is tolerant to the presence of artifact so that false positive alerts are reduced.
Upon sensing a life-threatening condition, a “caregiver”, who staffs the remote unit or “dispatcher station”, may remotely activate various devices, including an external defibrillator, pacer and drug infusion device, and/or may contact an EMS unit in the patient's immediate locale, in an attempt to save the patient's life.
The subject matter of both the U.S. Pat. Nos. 5,544,661 and 5,564,429 is incorporated herein by reference.
The systems disclosed in these two patents require that the portable patient unit be worn at all times so that ECG and other patient critical information can be continuously monitored. These systems cannot be used in a normal emergency situation where a patient has no advance warning of a cardiac event and is therefore neither prepared nor monitored by a lifesaving system. The U.S. Pat. No. 5,184,620 discloses a method of monitoring a patient's cardiac activity using a so-called “electrode pad” having a plurality of electrode sites which, upon placement against the patient's chest wall, provide ECG signals for determining if cardiac pacing and/or defibrillation is required. Certain combinations of electrodes provide the path for pacing signals whereas other combinations provide a path for defibrillation current.
The U.S. Pat. Nos. 5,593,426 and 5,782,878 (which have a similar disclosure) disclose a “communicator” for connecting each of a plurality of automatic external defibrillators (“AEDs”) to a central communication station. The central station receives information from an AED, such as patient ECG data and defibrillator operation data, and transmits information, such as use instructions for a bystander, to this AED.
The U.S. Pat. No. 4,102,332 describes a portable defibrillator, with a preprogrammed dialer, that telephones a physician when activated by a patient. While the physician and patient communicate with each other via the defibrillator's communication system, the physician can control the operation of the defibrillator from his or her remote location. During use, the defibrillator sends operation and status data to the physician.
The U.S. Pat. No. 6,141,584 discloses an automatic external defibrillator (AED) which is capable of storing ECG data and defibrillator data and “handing off” this data, via an infrared link, to equipment of emergency medical personnel when they arrive on the scene of a cardiac arrest.
The U.S. Pat. No. 6,148,233 discloses a wearable automatic external defibrillator; that is, a defibrillator which is worn by a patient having one or more contact electrodes attached to the chest wall of the patient for transmitting defibrillation energy to the patient and for receiving ECG information from the patient. This patent is directed specifically to the contact electrode(s) which can be worn by a patient for a relatively long time without skin irritation or damage. This system is designed for patients who have previously experienced cardiac arrhythmias but are, perhaps, not ready for an implanted defibrillator/pacer.
Finally, there are numerous patents which relate to ICDs. Such devices, which are also a form of automatic defibrillator, are in constant electrical communication with the human heart. When implanted, such ICDs operate independently, without external controls, to treat ventricular fibrillation (VF), ventricular tachycardia (VT) and supraventricular arrhythmias by applying one or more voltage pulses to the heart.
In their formal definitions, cardioversion refers to the delivery of a shock which is synchronized to the heart's electrical activity and defibrillation refers to an asynchronous shock. For simplification, unless otherwise noted, either one of the terms cardioversion and defibrillation shall hereinafter be referred to as simply defibrillation, and either one of the terms cardiovertor and defibrillator shall hereinafter be referred to as simply defibrillator. Such simplification is not intended to narrow the scope of the invention described herein, but is merely for the purpose of avoiding repeated use of the respective lengthy and rather awkward medical terminology.